AUTHOR=Xu Zhixiang , Ding Changsong 

TITLE=Combining convolutional attention mechanism and residual deformable Transformer for infarct segmentation from CT scans of acute ischemic stroke patients

JOURNAL=Frontiers in Neurology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2023.1178637

DOI=10.3389/fneur.2023.1178637

ISSN=1664-2295

ABSTRACT=Background: Segmentation and evaluation of infarct on medical images are essential for diagnosing and prognosis of acute ischemic stroke (AIS). Computer Tomography (CT) is the first-choice examination for patients with AIS.To accurately segment the infarct from the CT images of patients with AIS, we proposed an automated segmentation method combining the convolutional attention mechanism and residual Deformable Transformer in this paper. The method used the encoder-decoder structure, the encoders were employed for downsampling to obtain the feature of the images, and the decoder was used for upsampling and segmentation, respectively. In addition, we also further applied the convolutional attention mechanism and residual network structure to improve the effectiveness of feature extraction. Our code is available at https://github.com/XZhiXiang/AISsegmentation/tree/master.The proposed method was assessed on a public dataset containing 397 non-contrast CT (NCCT) of AIS patients (AISD dataset). The symptom onset to CT time was less than 24 hours.The experimental results illustrate that this work got a Dice coefficient (DC) of 58.66% for AIS infarct segmentation, which outperforms several existing methods. Furthermore, the volumetric analysis of the infarct indicated a strong correlation (Pearson correlation coefficient = 0.948) between the AIS infarct volume obtained by the proposed method and manual segmentation.The strong correlation between the infarct segmentation obtained from our method and the ground truth allows us to conclude that our method can accurately segment infarcts from NCCT.